Page 1 of 1

Lookup design

Posted: Sat Feb 07, 2015 11:32 am
by sam334
All,
Have a question in lookup design. we have a complex jobs which loads Table A everyday. Now, i want to make sure next day whatever record is there in Table A should not load. Used look up stage before loading into Table A.


--------------- Table A
------------------ |
Source-- Lookup- Table A.

Kind of confused in Lookup properties. In server jobs we use not(not found) in link variables in transformer. Seems like this is not available in Parallel.

Basically, suppose if 200 records comes from Source and it has look up based n employeeid, (let 1 record is already in Table A). it 199 record will load and 1 record will drop..

Any thoughts. Thanks.

Posted: Sat Feb 07, 2015 3:32 pm
by ray.wurlod
You have two choices in the Lookup stage.
  • Set the Lookup Failed property to Continue, making sure that all columns coming from the reference input are define as nullable there and on the output link, and check for NULL downstream.

    Set the Lookup Failed property to Output (the name of this value may change with different versions) and define a reject link from the Lookup stage.

Posted: Sat Feb 07, 2015 4:02 pm
by sam334
Okay. Thanks Ray. Awaiting for my membership activation though it is renewed almost 5 days back, it should be under process. As soon as it is done, will see your suggestion and try to implement. Thanks again.